Difference between revisions of "Fixing examples"
(→Project to fix the Einstein Toolkit example parameter files) |
|||
| (8 intermediate revisions by one other user not shown) | |||
| Line 9: | Line 9: | ||
I have submitted all arrangements/*/*/par/*.par files on 12 cores on datura, to get a first estimate of what happens. | I have submitted all arrangements/*/*/par/*.par files on 12 cores on datura, to get a first estimate of what happens. | ||
| − | + | === 14-Aug-2015 === | |
| − | |||
| − | |||
| − | |||
| − | + | The following tests take a very long time: | |
| + | |||
| + | * jpeg_amr | ||
| + | * jpeg | ||
| + | * http -> is supposed to take long. It starts an (empty) simulation and pauses it for user interaction at the web interface) | ||
| + | |||
| + | The current status of the examples is listed at [http://damiana2.aei.mpg.de/~ianhin/ETExamples/examples.xml Einstein Toolkit Examples Status]. This is regenerated manually at the moment. | ||
| + | |||
| + | Key: | ||
| + | |||
| + | {| class="wikitable" style="text-align:left;" | ||
| + | ! Status indicator | ||
| + | ! Description | ||
| + | ! No. of examples | ||
| + | |- | ||
| + | | Done | ||
| + | | Run ended with "Done." (TODO: eventually check that these are actually "good" somehow) | ||
| + | | 81 | ||
| + | |- | ||
| + | | ExternalThorns | ||
| + | | Failed due to depending on non-ET thorns | ||
| + | | 88 | ||
| + | |- | ||
| + | | Assertion | ||
| + | | Assertion failure | ||
| + | | 5 | ||
| + | |- | ||
| + | | CCTK_Abort | ||
| + | | Warning level 0 | ||
| + | | 43 | ||
| + | |- | ||
| + | | FailedParamCheck | ||
| + | | Failed Cactus parameter check | ||
| + | | 3 | ||
| + | |- | ||
| + | | Signal11 | ||
| + | | Segmentation fault | ||
| + | | 4 | ||
| + | |- | ||
| + | | Signal9 | ||
| + | | SIGKILL; probably out of memory, need to run on more nodes? | ||
| + | | 1 | ||
| + | |- | ||
| + | | Unknown | ||
| + | | Not categorised, but also includes runs which ran for a long time and were terminated manually for now | ||
| + | | 1 | ||
| + | |} | ||
| + | |||
| + | Proposals: | ||
| + | * Make sure that the examples all run in a resonable amount of time, e.g. less than one hour. Most of them are much quicker than this. For the long-running ones, see if it actually is necessary to have them take so long. Ideally they should run in a few seconds, unless there is a good reason. | ||
| + | * It would be good to have a list of the examples which can be run with just the ET on the web page, automatically generated, so that people can use these as the master list. We could also add a comment to the parfiles indicating whether they can be run using just the the ET. | ||
| + | * Fix the errors | ||
Latest revision as of 05:40, 14 August 2015
Project to fix the Einstein Toolkit example parameter files
See #641.
13-Aug-2015
(Ian)
I have submitted all arrangements/*/*/par/*.par files on 12 cores on datura, to get a first estimate of what happens.
14-Aug-2015
The following tests take a very long time:
- jpeg_amr
- jpeg
- http -> is supposed to take long. It starts an (empty) simulation and pauses it for user interaction at the web interface)
The current status of the examples is listed at Einstein Toolkit Examples Status. This is regenerated manually at the moment.
Key:
| Status indicator | Description | No. of examples |
|---|---|---|
| Done | Run ended with "Done." (TODO: eventually check that these are actually "good" somehow) | 81 |
| ExternalThorns | Failed due to depending on non-ET thorns | 88 |
| Assertion | Assertion failure | 5 |
| CCTK_Abort | Warning level 0 | 43 |
| FailedParamCheck | Failed Cactus parameter check | 3 |
| Signal11 | Segmentation fault | 4 |
| Signal9 | SIGKILL; probably out of memory, need to run on more nodes? | 1 |
| Unknown | Not categorised, but also includes runs which ran for a long time and were terminated manually for now | 1 |
Proposals:
- Make sure that the examples all run in a resonable amount of time, e.g. less than one hour. Most of them are much quicker than this. For the long-running ones, see if it actually is necessary to have them take so long. Ideally they should run in a few seconds, unless there is a good reason.
- It would be good to have a list of the examples which can be run with just the ET on the web page, automatically generated, so that people can use these as the master list. We could also add a comment to the parfiles indicating whether they can be run using just the the ET.
- Fix the errors
